Optimizing the CRISPR/Cas9 system for gene editing in Yarrowia lipolytica

Abstract

Yarrowia lipolytica is a promising host for producing valuable chemicals owing to its robustness and metabolic versatility. Efficient genome editing tools are essential for advancing its biotechnological applications. Although CRISPR/Cas9 technology has been applied in Y. lipolytica, achieving a consistently high editing performance remains challenging owing to the low homologous recombination efficiency and variability in system components. In this study, we optimized CRISPR/Cas9-mediated genome editing in Y. lipolytica to enhance its editing efficiency. Using the RNA polymerase III promoter SCR1-tRNA for sgRNA expression, we achieved a gene disruption efficiency of 92.5 %. The tRNA-sgRNA architecture enabled a dual gene disruption efficiency of 57.5 %. KU70 deletion in the Cas9 system increased the integration efficiency to 92.5 %, and Rad52 and Sae2 overexpression boosted homologous recombination. The introduction of Cas9^{D147Y, P411T} (iCas9) enhanced the efficiency of both gene disruption and genome integration. This study provides a powerful tool for efficient gene editing in Y. lipolytica, which will accelerate the construction of yeast cell factories.

Graphical abstract

1. Introduction

Yarrowia lipolytica is a non-conventional, oleaginous yeast recognized for its abilities to accumulate lipids and tolerate acidic environments, making it a valuable microbial cell factory for the production of various chemicals, such as organic acids [[1], [2], [3]], terpenoids [[4], [5], [6]], and lipids [7,8]. Advancements in the metabolic engineering of Y. lipolytica rely heavily on efficient genetic tools that enable precise genome modifications. The CRISPR/Cas system has emerged as a revolutionary genome editing tool owing to its simplicity, high efficiency, and versatility.

Applications of the CRISPR/Cas9 and CRISPR-Cas12a/Cpf1 systems in Y. lipolytica have been explored to enhance gene editing efficiency. Cas12 or Cpf1 uses a simple crRNA as a guide and recognizes TTTN, a protospacer adjacent motif (PAM) sequence. This unique PAM sequence may increase the on-target editing efficiency of Cpf1 owing to the lower chance of the enzyme misreading the sequence on genomes with a high GC content [9]. However, studies on Cas12 or Cpf1 for gene editing in Y. lipolytica are still limited [9,10]. Yang et al. [9] optimized the CRISPR/Cpf1 system and achieved high editing efficiency for counter-selectable markers and multiplexed genome editing in Y. lipolytica. Compared with Cpf1, the Cas9 system has been more widely used for gene editing in Y. lipolytica. Previous studies have focused on optimizing the individual components of the CRISPR/Cas9 system. Schwartz et al. [11] introduced CRISPR/Cas9 into Y. lipolytica, demonstrating that optimizing sgRNA expression using the SCR1′-tRNA^Gly promoter significantly improved the gene disruption efficiency of the system. Gao et al. [12] developed a single-plasmid system expressing both Cas9 and sgRNA, achieving notable efficiency in gene disruption and enabling the simultaneous disruption of two and three genes. Subsequent efforts by other researchers involved integrating Cas9 into the genome either by targeting specific loci such as the KU70 site [13] or through random insertion at other genomic sites [14], employing different promoters such as controlling sgRNA expression through the T7 promoter and T7 RNA polymerase [15], and modifying DNA repair pathways to inhibit non-homologous end joining (NHEJ) [13,16].

Despite these advances, achieving consistently high editing efficiency across different genes and experimental conditions remains challenging. The tendency of Y. lipolytica to repair DNA double-strand breaks (DSBs) via NHEJ rather than homologous recombination limits precise genome editing. Moreover, the variability in the strategies and conditions used in different studies renders comparison of the efficiency of the CRISPR/Cas9 system in Y. lipolytica difficult. Therefore, the systematic optimization of CRISPR/Cas9 expression strategies remains necessary. Moreover, several strategies that have been effective in enhancing the efficiency of the CRISPR/Cas9 system in other yeasts have not been fully explored in Y. lipolytica. For example, although Rad52 and Sae2 are known to improve homologous recombination efficiency [17], only Rad52 from Saccharomyces cerevisiae has been tested in Y. lipolytica and both of them have not yet been combined with the CRISPR system [18].

In this study, we systematically optimized the CRISPR/Cas9 system in Y. lipolytica by evaluating multiple factors influencing gene editing efficiency. We investigated the effects of sgRNA promoter selection, Cas9 expression and mutation, tRNA-sgRNA architectures, and DNA repair pathway modulation on the gene editing efficiency of the system. We compared the editing efficiencies of plasmid-based and genome-integrated Cas9 expression, offering valuable guidance for selecting gene-editing strategies on the basis of different requirements. In addition, we found that the overexpression of ScRad52 and ScSae2 (from S. cerevisiae) in Y. lipolytica improved the genome integration efficiency of the Cas9 system. Our work has enhanced genome editing efficiency in Y. lipolytica, facilitating its use in metabolic engineering and industrial biotechnology applications.

2. Materials and methods

2.1. Strains and media

Escherichia coli strain DH5α was used for plasmid construction and cultivated in Luria–Bertani medium (5 g/L yeast extract, 10 g/L peptone, and 10 g/L NaCl) supplemented with 0.2 mg/mL ampicillin. The solid medium was formed by the addition of 20 g/L agar. Y. lipolytica was cultivated in YPD medium (20 g/L glucose, 10 g/L yeast extract, and 20 g/L peptone) or synthetic complete (SC) medium (1.7 g/L yeast nitrogen base and 5 g/L ammonium sulfate, complemented with a corresponding complete supplement mixture).

2.2. Strains and plasmid construction

Y. lipolytica strains PO1f and PO1f-Cas9 [14] were used as chassis strains. In PO1f-Cas9, the TEFinp-Cas9-CYC1t expression cassette is integrated at the YALI1_E15321 g site [14]. Target DNA fragments were amplified using the polymerase chain reaction (PCR) with Phanta Max Super-Fidelity DNA Polymerase (Vazyme, Nanjing, China). Plasmids were constructed using the Gibson assembly method. The restriction enzymes were purchased from Thermo Fisher Scientific (Shanghai, China). All strains and plasmids used in this study are listed in the Supporting Information (Table S1).

2.3. Strain transformation and validation

The corresponding plasmids were transferred into the Y. lipolytica strains using the lithium acetate method [19]. For gene disruption, the transformants were cultured in an auxotrophic liquid medium specific to the introduced plasmid for 4 days, followed by streaking onto agar plates. Subsequently, 20 colonies were randomly screened using corresponding selection and YPD plates. Two plasmids—one carrying sgRNA or Cas9 and the other containing the integrated DNA fragment flanked by the homologous arms of the integration site—were transferred into the strains for gene integration. The transformants were cultured for 4 days, and 20 colonies were randomly selected. Subsequently, the genomes of the selected strains were extracted for PCR validation.

3. Results and discussion

3.1. Optimization of the CRISPR/Cas9 system to improve gene disruption efficiency

The level of sgRNA expression is a key factor in the gene editing efficiency of CRISPR systems. To optimize the expression of sgRNA, we compared different promoters to drive sgRNA expression and evaluated their effects on gene disruption efficiency. To evaluate the efficiency of the CRISPR system, we designed a guide RNA targeting the URA3 gene. DSBs were introduced and repaired by NHEJ, resulting in frameshift mutations that inactivated the gene. The efficiency of URA3 disruption was assessed by comparing the colony counts on SC-URA and YPD plates.

Previous studies have shown that both RNA polymerase II (Pol II) and RNA polymerase III (Pol III) promoters can drive sgRNA expression [11,20]. Pol II-derived RNAs typically contain a 5′ cap structure and a 3′ poly-A tail [20]. To generate functional Cas9/sgRNA ribonucleoproteins, the sgRNA must be flanked by hammerhead and hepatitis delta virus ribozymes to secure its release for activity [20] (Fig. 1A). By contrast, the Pol III promoters produce sgRNA directly without ribozyme processing (Fig. 1A). However, studies have demonstrated that using a Pol III promoter (e.g., SNR52) alone to drive sgRNA expression results in low gene editing efficiency [11]. To address this issue, Pol III promoters were combined with tRNA to enhance sgRNA expression [11]. Y. lipolytica naturally utilizes tRNA^Gly to transcribe 5S rRNA [21], and the tRNA maturation mechanism efficiently excises sgRNA from the primary transcript to produce the functional mature sgRNA [22].

Fig. 1.

Optimization of the CRISPR/Cas9 system for improved gene disruption efficiency. (A) Schematic of different promoters of sgRNA used in this study. sgRNA is a single guide RNA, HH is the hammerhead ribozyme, and HDV is the hepatitis delta virus ribozyme. Synthetic RNA polymerase (Pol) III promoters were generated by inserting the SNR52p or SCR1p sequences immediately upstream of tRNA^Gly. Polyt, which is a string of four thymines, served as the Pol III terminator. (B) URA3 disruption efficiencies achieved with different sgRNA promoters in the PO1f-Cas9-URA3 strain. (C) Validation of the efficiency of URA3 gene disruption. Transformants carrying different sgRNA promoters were randomly selected and streaked onto YPD and SC-URA plates, respectively. In total, 20 colonies from each biological parallel were picked and tested to calculate the gene disruption efficiency of the system. (D) URA3 disruption efficiency under different Cas9 expression conditions.

On the basis of these above-described research findings, we compared the performances of tRNA-combined Pol II and Pol III promoters in driving sgRNA expression. Using the Pol II promoter TEFin flanked by the hammerhead and hepatitis delta virus ribozymes, the efficiency of URA3 gene disruption reached 55 % (Fig. 1A–C). By comparison, when sgRNA expression was driven by the Pol III promoters SNR52-tRNA and SCR1-tRNA, the disruption efficiencies were 47.5 % and 92.5 %, respectively (Fig. 1A–C). This suggests that the SCR1-tRNA promoter is more effective in enhancing sgRNA expression and gene disruption in Y. lipolytica, as the Pol III promoter drives the efficient production of sgRNA while the tRNA stabilizes and accelerates its maturation, ensuring functional sgRNA for effective CRISPR editing [23].

Additionally, we compared the impacts of plasmid-based versus genome integration-based Cas9 expression on gene disruption efficiency. For plasmid expression, the sgRNA and Cas9 genes were inserted into a single episomal plasmid. The results indicated that when sgRNA was expressed under the control of the SCR1-tRNA promoter, the episomal plasmid expression of Cas9 achieved a relatively high gene disruption efficiency of 70 %, albeit it was still 22.5 % lower than that observed with genome-integrated Cas9 expression (Fig. 1D). Previous studies have shown that plasmid-based expression is not very stable [24,25]. We speculate that when the Cas9 gene is expressed from an episomal plasmid, its expression stability may affect its proportion of available protein and hence the editing efficiency.

3.2. Optimizing the incubation time to improve CRISPR efficiency

Although we achieved a high gene disruption efficiency, a prolonged incubation time in liquid medium under selection pressure was required after transformation of the yeast strains with the sgRNA-containing expression plasmid. Therefore, we investigated the effect of incubation time on gene disruption efficiency. After transformation of the PO1f-Cas9-URA3 strain with the sgRNA-carrying plasmid targeting URA3, we plated the transformants immediately (day 0), or after 2, 3, and 4 days of incubation in liquid medium under selection pressure, and tested the gene disruption efficiency of the system (Fig. 2A). The results showed that the CRISPR/Cas9 activity level was very low when the cells were plated immediately after transformation. The highest efficiency was achieved after 4 days of incubation, with a disruption efficiency of 82.5 %, which was slightly higher than that after 3 days (Fig. 2B). These results demonstrate that the duration of Y. lipolytica incubation in liquid medium following its CRISPR/Cas9 plasmid transformation significantly affects gene disruption efficiency. Extending the incubation period to 4 days maximized the Cas9-mediated gene disruption efficiency.

Fig. 2.

Comparison of the gene disruption efficiencies at different incubation times in selective liquid media after cell transformation with CRISPR/Cas9 plasmids. (A) Schematic representation of the transformation process and incubation in selective liquid media. (B) Gene disruption efficiency after different incubation times in selective liquid media.

3.3. Multiplex gene disruption using a tRNA-sgRNA architectural strategy

tRNAs play a crucial role in the transport of amino acids during protein synthesis. Composed of 70–90 nucleotides, tRNA sequences fold into cloverleaf-shaped structures. In eukaryotic cells, tRNA precursors are processed by RNase P and RNase Z, which cleave the redundant 5′ and 3′ sequences on the basis of the structural features of tRNA. When sgRNA is combined with a tRNA processing system, the high capacity of the endogenous tRNA processing system facilitates the simultaneous transcription of multiplex sgRNAs from a single promoter, thereby enabling efficient multiplex gene editing (Fig. 3A). Studies have shown that constructing sgRNA arrays using tRNA^Gly results in high gene editing efficiency [26].

Fig. 3.

Engineering of a tRNA-sgRNA architectural strategy for multiplex gene disruption. (A) Schematic diagram of multiplex gene disruption using the tRNA-sgRNA architectural strategy. (B) Efficiency of dual gene disruption when Cas9 is integrated into the genome. (C) Efficiency of dual gene disruption when Cas9 is expressed from an episomal plasmid. (D) Validation of the efficiency of URA3 and TRP1 disruption when a Cas9 expression cassette is integrated into the genome. The transformants were randomly selected and streaked onto YPD, SC-TRP, and SC-URA plates, respectively. In total, 20 colonies from each biological parallel were picked and tested to calculate the gene disruption efficiency of the system. (E) Validation of the efficiency of URA3 and TRP1 disruption when Cas9 is expressed from an episomal plasmid. The transformants were randomly selected and streaked onto YPD, SC-TRP, and SC-URA plates, respectively. In total, 20 colonies from each biological parallel were picked and tested to calculate the gene disruption efficiency of the system.

For dual gene disruption in Y. lipolytica, we constructed a YLEP-SCR1-tRNA-gURA3-gTRP1 plasmid to disrupt the TRP1 and URA3 genes simultaneously. Initially, we verified the effect of genome-integrated Cas9 expression on the efficiency of dual gene disruption. The efficiencies of single gene disruption were 62.5 % and 95 % for TRP1 and URA3, respectively (Fig. 3B and 3D). Notably, the efficiency of simultaneous TRP1 and URA3 disruption was 57.5 % (Fig. 3B and 3D). We also tested the efficiency of dual gene disruption using an episomal plasmid expressing the Cas9 gene. The pCAS1yl-SCR1-tRNA-gURA3-gTRP1 plasmid was introduced into the PO1f-URA3 strain to verify the efficiency of URA3 and TRP1 disruption. The dual gene disruption efficiency was only 10 %, which was 47.5 % lower than that achieved when Cas9 was integrated into the genome (Fig. 3C and 3E). The results demonstrate that genome-integrated Cas9 expression significantly outperforms plasmid-based systems for dual gene disruption in Y. lipolytica. This highlights the importance of stable Cas9 expression, as the lower efficiency in the plasmid-based system is likely due to the reduced expression or stability of the Cas9.

3.4. Enhancing CRISPR-mediated genome integration efficiency

Harnessing the CRISPR system to integrate DNA into the genome without a selectable marker can facilitate rapid strain development. In addition to the sgRNA targeting and Cas9 cleavage capacities of the CRISPR system, its homologous recombination capability is also an important factor for genome integration. Y. lipolytica has a strong NHEJ preference but poor homologous recombination efficiency (Fig. 4A).

Fig. 4.

Enhancement of the CRISPR-mediated genome integration efficiency. (A) Schematic representation of the CRISPR-mediated precise repair of double-strand breaks for gene integration. (B) Schematic representation of the plasmid used for gene integration with genome-integrated Cas9 expression. (C) Gene integration efficiency of genome-integrated Cas9 upon the disruption of KU70 or overexpression of YlRad52, ScRad52, or ScRad52-ScSae2. (D) Schematic representation of the plasmid used for gene integration with plasmid-based Cas9 expression. (E) Gene integration efficiency of episomal plasmid-borne Cas9 upon KU70 disruption.

Previous studies have shown that deletion of the KU70 gene can reduce NHEJ activity [12,13], whereas overexpression of Rad52 or Sae2 can enhance homologous recombination efficiency [17,18]. However, the combination of Rad52 and Sae2 with the CRISPR system has not been tested in Y. lipolytica. First, we tested the efficiency of gene integration under genome-integrated Cas9 expression. To assess the integration efficiency, we co-transformed yeast cells with two plasmids: one containing the sgRNA sequence and the other containing a GFP expression cassette flanked by 1 kb homology arms (Fig. 4B). The integration efficiency after KU70 deletion increased to 92.5 %, showing a substantial improvement over that of the wild-type control (PO1f-Cas9), indicating that disruption of the KU70 gene is an effective strategy for boosting homologous recombination-mediated genome integration (Fig. 4C). We also investigated the effects of overexpressing homologous recombination-related genes from different sources (Rad52 from Y. lipolytica, Rad52 from S. cerevisiae, and Sae2 from S. cerevisiae). YlRad52 overexpression improved the integration efficiency by 7.5 %, whereas ScRad52 overexpression was more effective, achieving a 20 % increase over that of the control. However, the combined overexpression of ScRad52 and ScSae2 did not further enhance the integration efficiency of the system (Fig. 4C).

We also tested the efficiency of gene integration when Cas9 was expressed from an episomal plasmid. To assess integration efficiency, we co-transformed the yeast cells with two plasmids: one containing both the Cas9 expression cassette and sgRNA and the other containing the targeted gene flanked by 1 kb homology arms (Fig. 4D). When wild type Cas9-containing plasmid was transformed in PO1f strain, the integration efficiency was 42.5 % (Fig. 4E). KU70 disruption slightly increased the integration efficiency, which reached 47.5 % (Fig. 4E). Plasmid-based Cas9 expression resulted in significantly lower gene integration efficiency than that achieved with genome-integrated Cas9 expression. This may be due to the instability of plasmid-based Cas9 expression, leading to insufficient DSBs to fully activate the homologous recombination repair pathway.

3.5. Enhancing gene editing efficiency using Cas9^{D147Y, P411T} mutations

Compared with genome-integrated Cas9 expression, episomal plasmid-based Cas9 expression resulted in significantly lower efficiency in both multiplex gene disruption and gene integration. To address this issue, we sought to enhance the gene editing efficiency. Previous studies have shown that Cas9^{D147Y, P411T} (iCas9), a Cas9 mutant, is crucial for improving gene editing efficiency [27]. iCas9 harbors two amino acid mutations: D147Y and P411T. The D147Y mutation eliminates the negative charge on the Cas9 protein surface, thereby optimizing the surface charge of the protein and enhancing its specificity for target DNA binding [28]. The P411T mutation alleviates the inhibitory effect of the cyclic structure of Pro-412 on sgRNA binding, thereby increasing the cleavage efficiency and expanding PAM compatibility [28]. Using plasmid-borne iCas9, the efficiency of single URA3 disruption reached 100 % and that of dual disruption of URA3 and TRP1 reached 22.5 %, which were respectively 77.5 % and 12.5 % increases over the efficiency achieved with wild-type Cas9 (Fig. 5A and 5B). We also evaluated gene integration efficiency by co-transforming the yeast cells with two plasmids: one carrying the iCas9 expression cassette and sgRNA and the other containing the target gene flanked by 1 kb homology arms (Fig. 5C). Expression of iCas9 in PO1f increased the integration efficiency to 75 % (Fig. 5D), demonstrating that iCas9 maintains high cleavage efficiency even at low expression levels, which will facilitate genome engineering in Y. lipolytica. We also investigated the effects of KU70 disruption, ScRad52 overexpression, and ScRad52-ScSae2 co-overexpression on this system. However, these did not lead to further improvements in gene integration efficiency (Fig. 5D).

Fig. 5.

Improvement of the gene editing efficiency with plasmid-borne iCas9. (A) Efficiency of dual gene disruption achieved with episomal plasmid-based iCas9 expression. (B) Validation of the efficiency of URA3 and TRP1 disruption when iCas9 is expressed from an episomal plasmid. The transformants were randomly selected and streaked on YPD, SC-TRP, and SC-URA plates, respectively. In total, 20 colonies from each biological parallel were picked and tested to calculate the gene disruption efficiency of the system. (C) Schematic representation of the plasmid used for gene integration with plasmid-based iCas9 expression. (D) Gene integration efficiency achieved with episomal plasmid-borne iCas9 and iCas9 in combination with KU70 disruption, ScRad52 overexpression, or ScRad52-ScSae2 co-overexpression.

Additionally, we explored the effect of the donor DNA conformation on the efficiency of CRISPR-mediated integration. A linear DNA fragment (UT8p-hrGFP-CYC1t) and plasmid-borne DNA were tested for site-specific integration at the B3 locus (Fig. S1A). The results revealed that when the donor DNA was introduced as a linear fragment, the integration efficiency was only 2.5 % (Fig. S1B). By contrast, the efficiency of integration of the plasmid-borne DNA increased by 57.5 % (Fig. S1B). The higher integration efficiency of plasmid-borne DNA may be due to its greater stability and reduced susceptibility to degradation, allowing for a more sustained supply of donor DNA.

4. Conclusion

In summary, we have systematically optimized CRISPR systems in Y. lipolytica. We found that Pol III promoters, particularly SCR1p-tRNA, significantly enhanced the sgRNA expression level and gene disruption efficiency compared with Pol II promoters. The tRNA-sgRNA architectural strategy achieved dual gene disruption with an efficiency of 57.5 % in Y. lipolytica. When a Cas9 expression cassette was integrated into the genome, the gene integration efficiency reached 92.5 %. With the episomal plasmid-based expression of iCas9 (D147Y, P411T), the gene integration efficiency increased to 75 %. These optimized CRISPR systems will facilitate the precise manipulation of genomes to construct efficient microbial cell factories.

Data Availability Statement

All relevant data supporting the findings of this study are available in this manuscript and the supplementary materials.

CRediT authorship contribution statement

Jianhui Liu: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Yamin Zhu: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Jin Hou: Writing – review & editing, Writing – original draft, Resources, Project administration, Funding acquisition.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.